ELSEVIER MaterialsScience and Engineering: C 3 (1995) 125-130

Biomimicry ofbamboo bast fiber with engineering composite materials

S.H. Li a,b,*, Q.Y. Zeng , Y.L. Xiao a, S.Y. Fu a*b,B.L. Zhou a,b
aInternational Centerfor Materials Physics, Academia Sinica, 72 Wenhua Road, Shenyang 110015, Peoples Republic of China
b Institute qf Metal Research, Academia Sinica, 72 Wenhua Road, Shenyang 110015, Peoples Republic of China
Institute of AppliedEcology, Academia Sinica, 72 Wenhua Road, Shenyang llOOl5, Peoples Republic of China

Abstract

Bamboo, one of the strongest natural structural composite materials, has many distinguishing features. It has been found that its reinforcement
unit, hollow, multilayered and spirally-wound bast fiber, plays an extremely important role in its mechanical behavior. In the present work,
on the basis of the study on bamboo bast fiber and wood tracheid, a biomimetic model of the reinforcing element, composed of two layers of
helically wound fiber, was suggested. To detect the structural characteristics of such a microstructure, four types of macro fiber specimens
made of engineering compositses were employed: axially aligned solid and hollow cylinders, and single- and double-helical hollow cylinders.
These specimens were subjected to several possible loadings, and the experimental results reveal that only the double-helical structural unit
possesses the optimum comprehensive mechanical properties. An interlaminar transition zone model imitating bamboo bast fiber was proposed
and was verified by engineering composite materials. In our work, the transition zone can increase the interlaminar shear strength of the
composite materials by about 15%. These biomimetic structural models can be applied in the design and manufacture of engineering composite
materials.

Keywords: Bamboo; Bast fiber; Biomimetics;Engineering composites

1. Introduction Bamboo is another kind of common natural structural com-

Mankind has long had the idea that we should learn from
nature. There are innumerable biological materials in nature
that exhibit distinguishing performance. All biological mate-
rials are, without exception, Icomposites. With therapiddevel-
opment of engineering composite materials in recent years,
various problems were encountered in both the design and
fabrication process. To solve these newly emerging problems,
many measures were attempted. Among them, one interesting
method became increasingly attractive; this new interdisci-
plinary subject is called biomimetics [ 1,2], the purpose of
which is to advance man-made composite materials through
guidance from nature. In the former biomimetic study, wood
as the most common natural: composite, was noticed as early
as 1980. Gordon and Jeronimidis extensively studied the tra-
cheid of timber, and successfully applied the energy absorp-
tion mechanism of wood cells under tension in engineering
composite materials to increase the work of fracture [ 31. The
thorough understanding of the interplay between the mor-
phology and structural performance of different types of
wood led to a remarkable (development of a patented new
material [ 41.
* Correspondingauthor.
posite. It is a widespread plant family found in all continents.
Apart from being a basic material in the construction and
paper industries, bamboo is a more typical long-fiber rein-
forced composite than wood. Thus, some biomimetic studies
on bamboo have been carried out in recent years by Li et al.
[ 51. Earlier works were primarily focused on the macro and
meso scales. From the results of the present work, it was
found that on a micro scale, bamboo has some unique features
which are essentially important for the high performance of
bamboo.
2. Comparison of wood and bamboo
Wood is a natural composite that exhibits a remarkable
combination of strength, stiffness and toughness. Detailed
studies have established that the unique hierarchical architec-
ture with which the constituents of wood are arranged is the
basis for achieving excellent properties. Besides wood, bam-
boo has a gradient structure and excellent specific properties
which are even comparable with ARALL@ [ 51. The chemical
composition of wood and bamboo [ 61 are listed in Table 1.
From Table 1 it can be seen that bamboo and wood have
very similar chemical components, but their mechanical

09284931/95/$09.50 0 1995Elsevier Science S.A. All rights reserved

SSDIO928-4931(95)00115-8
126 S.H. Li et al. /Materials Science and Engineering: C 3 (1995) 125-130

Table 1 features of the reinforcing elements of biomaterials were

Chemical composition and tensile strength of wood and bamboo taken into consideration and were experimentally investi-
Wood Bamboo [6]
gated under all possible loading types. The modulus proper-
ties of bamboo bast fiber were reported in Ref. [ 81, and only
Cellulose ( %) 40-50 45.3 the strength properties were focused on.
Hemi-cell;lose (%) 20-35 From Fig. 1 it can be seen that a bamboo fiber consists of
Lignin (%) 15-35 25.5
alternating broad and thin concentric layers which are com-
Polyoses (%) 24.3
Extractive (I) < 10 2.6
posed of microfibrils. The microfibril orientation angles in
Tensile strength ( MPa) 34-220 150-520 the broad layer are in the range 3-10 with respect to the fiber
axis, and those in thin layers are in the range 30-90, but
mostly 30-45. Although its microstructure is complicated,
we assume that a broad and a thin layer form a structural unit.
L4 Imitating such a structural unit, a biomimetic double-helical
N3 reinforcing element model, composed of a broad and a thin
L3 layer was proposed.
N2 Corresponding to all the structural features mentioned
L2
above, four types of macrofiber specimens were designed and
manufactured in order to reveal the structural characteristics.
NI
The material is glass fiber reinforced epoxy resin. Although
LI these fiber specimens are on a macro scale, it is reasonable to
0 treat them as the result of magnifying the microfibril and
P bamboo bast fiber in the same proportion, and hence the
structural and mechrinic&l features of the microscopic bast
fiber will still be maintained. The structures of the four types
of macrofiber are shown in Fig. 2. Their structural forms are
Fig.1.Schematicdiagramof woodtracheid(a) and bamboo bast fiber (b)
Primary wall (P), outermost layer of the second wall (0). broad layers listed in Table 2. The total fiber number and fiber volume
(Ll-LA), and narrow layers (Nl-N3). (After Wai et al. [7].) fraction of each type of specimen were controlled to be the
same in the manufacturing process in order to make the com-
properties are very different. Their structural differences, parison of the test results meaningful. The properties tested
from macro to micro scale, are responsible for this. On the
macro scale, wood is a solid cylinder, while bamboo is a
hollow cylinder with many nodes arranged longitudinally.
On the meso scale, wood is composed of alternative spring
and summer wood, but bamboo has a non-linear gradient
structure which comprises vascular bundles and thin-walled
cells. On the micro scale, wood tracheid and bamboo bast
fiber are both hollow tubes or cylinders composed of several
concentric layers and each layer is reinforced with helically
wound microfibrils (or protofibrils). However, the differ-
ences in their cell wall thickness, layer number and microfi-
brillar angle are also obvious. In Fig. 1 the microstructures
of a wood cell (a) and a bamboo fiber (b) [7] are shown.
(a) (b) (c) (d)

3. Biomimetic study of a double-helical reinforcing

element model
The common structural features of the reinforcing elements
of plant materials can be expressed as being all hollow, hel-
ically wound and multi-layered cylinders, or tubes. In pre-
vious biomimetic research, only the hollow and helically
wound features were noted and studied [3,4], and the way
of loading was only limited to tension. In fact, the reinforcing
elements of living organisms are possibly subjected to other
kinds of loads, such as compression, bending, etc., even a
combined complicated load. In this work; all the structural
Fig. 2. Diagram of the structures of the four types of macrofiber specimens.
(a) Solid cylinder with the fibers axially aligned. (b) Hollow cylinder with
the fibers axially aligned. (c) Hollow, single-helical cylinder. (d) Hollow,
double-helical cylinder.
Table 2
Four types of macrofibers to verify the structural features of bamboo fibers
A: Solid cylinder, its fibers are arranged along the axial direction
B: Hollow cylinder, its fibers am arranged axially
C: Hollow single-layered right-helical cylinder (15)
D: Hollow double-layered helical cylinder (80% fibre inside in right helix,
15;20% fibre outside in left helix, 30)
 S.H. Li et al. /Materials Science and Engineering: C 3 (1995) 125-130 127

Table 3
The properties tested and geometric size of the specimen (D and dare the outer and inner diameters of the macro fiber specimen, respectively)

Properties tested Specimen length Specimen diameter

(mm) (mm)

Compressive buckling strength 300 A: D= 13.8

Three-point bending 230 (span= 190) B,C&D: 0=14.7;d=5
Compressive strength 30
Post-buckling compressive strength 20
Tensile modulus and strength 210 A: D=7.5
Gauge length= 80 B,C&D:D=8;d=2.8

Table 4
A summary of the experimental :results on the six mechanical properties of the four types of structures (the data in parentheses are standard deviations)

No. of A B C D
samples

Compressive buckling Vr=40% 49.4 - 38.9 49.4

strength ( MPa) (5.1) (7.4) (4.7)
v,= 50% 63.9 89.5 53.1 82.5
(6.8) (14.1) (8.3). (5.9)
Compressive strength V,=40% 138.1 - 128.5 141.2
(MPa) (8.4) (7.3) (5.9)
v*= 50% 201.1 262.1 199.3 208.1
(20.4) (29.9) (24.1) (18.9)
Apparent flexural strength 251.7 262.2 253.0 253.9
(MPa) (14.2) (21.8) (19.2) (16.1)
Post-buckling compressive 207.5 213.0 180.6 176.8
strength (MPa) (32.2) (29.1) (33.4) (17.4)
Tensile mo,dulus ( GPa) 25.6 22.1 22.1
(2.2) (1.8) (1.4)
Tensile strength ( MPa) 410.7 373.9 401.3
(24.6) (16.2) (21.2)

and the geometric size of the specimens are listed in Table 3.
The mechanical properties tested and the experimental results
are summarized in Table 4.
Natural bamboo bast fiber is so slender that its diameter
ranges between 20 pm and 45 pm and its length is lOOO-
2000 pm. For a single fiber, the stability of such a structure
with a high length to diameter ratio cannot be overlooked.
Thus the axial compressive buckling test was performed to
compare the structural stability of the four types of specimens.
The testing speed was 2 mm min- and the test was per-
formed at room temperature. From the experimental data
listed in Table 4, it can be seen that the D-type specimen is
more stable than the C-type specimen but less than the B-
type, and shows no inferiority to the A-type specimen. That
is to say, double-helical structures will be more stable than
single-helical ones, although only a thin layer of left-helical
fibers was wound outside. The inversely wound helical fiber
layers restrained each other when the specimen was com-
pressed During the test for axial compressive buckling, an
interesting phenomenon was observed. Of the four types of
specimens, only on the surface of the D-type specimen was
obvious damage found after buckling. This can be explained
as follows. When buckling occurs, the specimen will bend in
the central part and hence relative displacement occurs
between two oppositely arranged helical layers and the out-
side thin layer is more easily damaged. To determine the
extent of the damage caused by buckling, the post-buckling
compressive strength was tested and the results are also listed
in Table 4. From these data it is evident that buckling causes
more damage to double-helical structures than the other three
types of structure. This is also the reason why D-type struc-
tures can improve the structural stability.
For the bending test, owing to the anisotropy and thick
wall of the four types of specimens, the precise calculation
of the maximum normal stress according to classical com-
posites theory is very complicated. Here the geometric sizes,
fiber volume fraction and total number of fibers were con-
trolled as seriously as possible in the manufacturing process
of the specimens so that a comparison of the maximum bend-
ing load is meaningful. Furthermore, for the A-type speci-
men, its fibers were arranged along the axial direction, and
the deviation angle of the C-type specimen was. 1.5. For the
D-type specimen, there are only 20% of fibers whose helical
angle is 30, thus the relative anisotropy is not very strong. If
the anisotropy of the specimen was not taken into account,
the maximum normal stress, calculated according to the
128 S.H. Li et al. /Materials Science and Engineering: C 3 (1995) 125-130
40
30
s
0 20
(P
s
10
0
0 I 2 3 4 5
Dirplacement (mm)
Fig. 3. Typical curve of the D-type specimen under compressive load.
mechanics of the materials, can give a rough comparison of
the maximum normal stress. This can be called the apparent
flexural strength, a parameter of the relative capacity to bear
a bending load. From the calculated results listed in Table 4,
it can be seen that the D-type specimens show no inferiority
to the other types of structure in bearing the bending load.
For the compressive test, the load-displacement curve of
the D-type specimen is much different from those of the
others. Its typical curve is shown in Fig. 3. If only from the
viewpoint of compressive strength, the D-type structure is
closer to the A type, but is better than the C-type structure.
However, if the pseudo-plastic phenomenon which occurred
in the D-type specimen is considered, the structural advantage
of the D-type specimen is evident. The displacement of the
D-type structure is increased as much as twice, but its strength
is maintained at a level of 85% of the maximum compressive
strength. In other words, at the expense of 15% strength loss,
the plasticity of the double-helical structure is increased by
200%.
The tensile property is always very important for any struc-
ture. To compare the tensile behavior of the spiral and non-
spiral structures, three types of tensile specimens were made
and tested. Since the tensile load is very high and the pressure
stress of the tabbed region is of considerable value, to prevent
a collapse occurring in the end part under the pressure of jigs,
the tabbed region was filled inside with iron wire and was
reinforced outside with glass fiber and epoxy resin. The test-
ing speed for the tensile modulus is 2 mm min- and 5 mm
min- for the tensile strength. The results of the apparent
tensile modulus and the tensile strength of the three types of
specimen are listed in Table 4. For the tensile property,
according to the classical theory of fibrous composite mate-
rials, there is no difference between structures A and B.
For the spirally wound tube under tension load, an inter-
esting phenomenon of tension buckling was observed by Page
et al. when they studied wood tracheid in 197 1 [ 91. Although
not all the fibers in a given cross-section of wood actually
buckle in this manner when wood fractures, a considerable
Dieplacement (mm)
Fig.4. Load-displacement curve of the type-C specimen under tension.
number of them, perhaps about lo%, do. This phenomenon
was then applied in engineering composite materials and the
work of fracture was increased successfully at the expense of
a moderate loss of stiffness and axial strength [ 31.
In our tension experiment, the so-called pseudo-plasticity
existed only in the C-type specimen, as shown in Fig. 4,
although not as obvious as reported in Ref. [ 31 for the long
spirally wound tube. The fiber fully aligned A-type specimens
fracture abruptly after the maximum load. This can be
explained as follows. According to the buckling theory of
composite materials, the critical buckling stress is a function
of the principal elastic constants of the tube, the angle of the
spiral winding, and tube thickness [ lo]. Compared with the
specimen in Ref. [ 31, the specimens in our work have thicker
tube walls. For the D-type specimen, the double-helical ele-
ment, the two inversely wound fiber layers restrain each other
and thus reduce the possibility of buckling. In other words,
the structural stability is increased. However, compared with
the plain A-type specimen, the double-helical structure
improves the plasticity to some extent.
What was analyzed above is for a single fiber cylinder.
However, compared with the A-type structure, which can be
thought of as a model of a bundle of fibers in laminates, the
D-type structures have other advantages when they are used
in bundle form. The firm attachment of a fiber cylinder to
other adjacent ones can effectively prevent them from twist-
ing. Hence, there exists a restraint against any shear strain
which might otherwise result from the application of a normal
stress along the fiber cylinder axis. More specifically, the
restraint arises because, in the double cylinder wall of two
adjacent cylinders, the same helical sense of equivalent layers
in both cylinders will mean opposite signs of the helical angle
in the two layers. The shear strains of the separate equivalent
layers would thus also be of opposite sign, but since the layers
are connected and two cylinder walls must deform as a unit,
the shear strain is effectively restrained.
On the basis of above experimental data and analysis, a
useful conclusion can be drawn: a double-helical fiber struc-
 S.H. Li et al. /Materials
ture has the optimum comprehensive performance compared
with other structural forms. A single-helical fiber structural
model is too simple to reproduce the superiority of a plant
fiber which is essentially a multi-layered helical structure.
The double-helical structural element--one thick layer with
a small spiral microfibrillar angle and a thin layer with a
greater angle-is probably the structural unit of bamboo bast
fiber. For its engineering application, the thickness of two
layers and the spiral angle of the fiber in it can be designed
to meet different requirements for plasticity, strength and
structural stability.
4. Biomimicry of the transition zone in a bamboo fiber
We know that bamboo bast fiber possesses a very compli-
cated structure, from the macro to micro scale. In particular,
bamboo bast fiber is, to some extent, similar to other kinds
of plant cells or fibers, as we described earlier. However, the
ultrafine structure of bamboo bast fiber is somewhat different
from others, at least from the published literature. From the
observations of Wai et al. [7], bamboo bast fiber has a
detailed ultrafine structure between broad and thin layers, as
shown in Fig. 5 (b) .
From Fig. 5 it can be .seen that bamboo fiber contains
alternating broad and thin layers. Let us pay special attention
to the interface between the thick and thin layers, namely the
transition zone. Usually there are several sub-layers in the
transition zone and the microfibrils there change their eleva-
tion angle gradually from one layer to the adjacent one. This
phenomenon undoubtedly increases the complexity of the
microstructure of a bamboo fiber. From the viewpoint of
biological evolution they must have some function. What is
the use of the transition zone? Obviously the continuity of
the physical properties of the adjacent layers can be improved.
Do they have any influence on the mechanical behavior of
bamboo fiber?
It is natural for us to i-magine that this transition zone
probably has some effect on the inter-layer property. So, a
L4
N3
L3
- N2
- LI
(a) (b)
Fig. 5. Schematic illustration of the secondary wall stracture of a bamboo
fiber. (a) The struch~re of a fiber; (b) the detailed structure of (a) showing
the transition zone. (After Wai et al. [ 71.)
Science and Engineering: C 3 (1995) 125430 129
Fig. 6. A simplified biomimetic model of the transition zone imitating bam-
boo fiber.
simplified plane model rather than a real spatial spiral one
was designed (as shown in Fig. 6) and was verified by engi-
neering composite materials.
Between two normal layers, a transition zone was formed
by inserting several additional sub-layers in order to avoid
the abrupt change in the angle of the principal axis of the two
adjacent layers. The sub-layers are thinner than the normal
ones. The fiber angle of the sub-layer was designed in such a
way that the sub-layers will divide the fiber angle difference
of the adjacent layer into equal parts. First, an engineering
composite material of glass fiber reinforced epoxy resin was
employed, and for comparison another kind of specimen
without a transition zone was also prepared (like Fig. 6).
For many engineering applications, composite materials
are used in laminated form and in such cases delamination is
one of the main modes of damage and failure. The interlam-
inar shear modulus and ultimate stress are therefore both very
important parameters to be considered in the design of struc-
tures. Several methods have been suggested to measure the
interlaminar shear strength of fiber-reinforced composites.
The most commonly used method is to subject a short thick
beam to three-point loading until failure occurs by shear on
the central plane. It has also been reported [ 111 that the short
beam method cannot be recommended as an accurate way to
acquire data for design purposes. Besides short beam, the
double-side grooved specimen [ 121 becomes more common
in the test of interlaminar shear properties.
Here, both short beam specimens and grooved specimens
were used to measure the interlaminar shear strength (ILSS) .
For the glass fiber/epoxy resin specimens, the interlaminar
L3 shear strength was measured by using a two-sided grooved
specimen, as shown in Fig. 7. The experimental results are
shown in Fig. 8, where it can be seen that the transition zone
N2 slightly increased the interlaminar shear strength.
To verify this conclusion, another engineering composite
material was also used, i.e. glass fiber reinforced polyester.
This time the ply mode was [ + 12/- 12],, and ( + 12),/
+4/ - 4/ ( - 12),. The sub-layer with the angle of 4 is made
from different glass strands and is thus thinner. This time the
short beam specimen for three-point bending was exploited
to test interlaminar shear strength. Eight specimens were
tested for each group and the results are shown in Fig. 8. The
span is 19 mm and the diameter of the nose is 20 mm.
130 S.H. Li et al. /Materials Science and Engineering: C 3 (1995) 125-130
m Without B with
tranrition zone tranrltion zone
a
2 kle and elongate. It was reported that in wood only a small
6o 3
2 50 .
F
s 40 -
(10
$ 30.
f 20 -
a
.E 10 .
ii
i:0 O-
E GF/Epoxy QF/Polyerter
Fig. 8. Verification of the biomimetic model of the transition zone with glass
fiber/epoxy resin and glass fiber/polyester.
From Fig. 8, the ILSS of the specimens with a transition
zone are about 15.8% and 13.3% higher than those without
for glass fiber/epoxy and glass fiber/polyester, respectively.
On the basis of the above experimental results, we can draw
the following conclusion: the sub-layer in the transition zone
of bamboo will increase the inter-layer shear property and
improve the geometrical and physical continuities of the adja-
cent layers.
The biomimicry experiments above have a two-fold mean-
ing: on the one hand, they give solid proof that the transition
zones in bamboo fiber have a significant effect on the inter-
layer property. On the other hand, the concept stored in nat-
ural biomaterials can be applied in engineering composite
materials, especially where the interlaminar property is of
particular importance. It can easily be inferred that the tran-
sition zone will also increase the interlaminar shear modulus
within the area between the two principal axes of the adjacent
layers.
5. Some discussion about the application of these
biomimetic models in engineering composite materials
The double-helical reinforcing element model can be
applied to engineering composite materials for better com-
prehensive mechanical behavior. To meet the different requi-
rements, the thickness ratio of the broad and thin layers, the
elevation angle of the fiber in each layer, and the number of
reinforcements with pseudo-plasticity can all be designed. To
improve the plasticity, some amount of single-helical rein-
forcements could be combined with double-helical ones, so
that under certain tensile stress, some reinforcement will buc-
amount of the fibers ( 10%) will buckle when wood fractures.
For the biomimetic transition zone model, it can easily be
applied in laminates to improve the interlaminar shear prop-
erties.
6. Conclusions
( 1) The double-helical structural unit possesses the opti-
mum comprehensive mechanical properties, The single-hel-
ical structure is too simple to reproduce the superiority of
plant fibers which are essentially multi-layered helical struc-
tures.
(2) The transition zone in bamboo bast fiber can improve
the inter-layer property. The verification tests show that the
interlaminar shear strength of glass fiber/epoxy resin and
glass fiber/polyester are increased by about 15% after the
introduction of such a transition zone.
Acknowledgment
This work was performed under the auspices of the
National Nature Science Foundation of China.
References
[ 1] A.V. Srinivasan, G.K. Haritos and F.L. Hedberg, Appl. Mech. Rev., 44
(1991) 463.
[2] J.F.V. Vincent and A.V. Srinivasan, Biomimetics, I (1992) 1.
[3] J.E. Gordon and G. Jeronimidis, Phil. Tram. R. Sot. Lon., A 294
(1980) 545.
[4] R.C. Chaplin, J.E. Gordon and G. Jeronimidis, U.S. Patent no.4, 409
(1983) 274.
[5] S.H. Li, Q.Y. Zeng, S.Y. Fu and B.L. Zhou, &mime&, 2 (1993)
15.
[6] D. Fengel and X. Shao, Wood&i. Technol., 18 (1984) 103.
[7] N.N. Wai. H. Nanko and K. Murakami, Wood.%. Technof., 19 ( 1985)
211.
[8] S.H. Li, Q.Y. Zeng, X. Chen and B.L. Zhou, J. Mater. Sci. Technol.,
10 (1994) 34.
[9] D.H. Page, F.El-Hossseiny and K. Winkler, Nature, 229 (1971) 252.
[ 101 N.J. Pagano, J.C. Halpinand J.M. Whitney,.I. Camp. Mater., 2 (1968)
154.
[ 111 J.B. Sturgeon, RAE Technical Reporr 71026 (1971).
[ 121 M.F. Markhan and D.Dawson, Composires, 6 (1975) 173.